Development of a scoring system for predicting adjacent vertebral fracture after balloon kyphoplasty

BACKGROUND CONTEXT The incidence of adjacent vertebral fracture (AVFs) is reported to be 10%–38% after balloon kyphoplasty. However, no reports have established a system for prediction of AVF occurrence. PURPOSE To establish a scoring system for predicting AVF occurrence after balloon kyphoplasty for osteoporotic vertebral fractures (OVFs). DESIGN A prospective cohort study. PATIENT SAMPLE Consecutive elderly patients aged 65 years and older who underwent balloon kyphoplasty for OVFs within 2 months after the onset. OUTCOME MEASURES AVF was confirmed by X-ray. METHODS From 2015 to 2017, 116 consecutive patients from 10 participating hospitals who underwent balloon kyphoplasty were enrolled in this study. Prior to study enrollment, each patient underwent plain X-ray, computed tomography (CT), and magnetic resonance imaging (MRI) of the thoracic and lumbar spine. Severity of pain was subjectively assessed using a visual analog scale (VAS) based on the average level of back pain that the patient had experienced in the preceding week. After enrollment, subjects underwent balloon kyphoplasty. Quality of life was evaluated using SF-36. Patients were followed up for at least 6 months. RESULTS Of the 116 patients enrolled, 109 patients with all the required data at the time of enrolment and the 6-month follow-up were included in the study. A total of 32 patients (29%) showed AVFs within the 6-month follow-up. No significant differences were observed in each clinical outcome at 6-month follow-up, although higher VAS score for back pain at 1-month follow-up was observed in the AVF group (37.5) than in the non-AVF group (20.8, p<.001). Wedge angle of vertebrae before surgery was greater in the AVF group (21.6°) than in the non-AVF group (15.7°, p<.001). The change in wedge angle between pre- and postsurgery was greater in the AVF group than in the non-AVF group, whereas the change in local kyphosis was not significantly different. The multiple logistic regression model showed increased odds ratio (OR) of thoracic or thoracolumbar spine, old OVF presence, >25° kyphosis before surgery, and >10°correction for AVF. Based on this result, a simple scoring system for predicting AVF occurrence was developed. The total AVF score was calculated as the sum of the individual scores, which varied from 0 to 6. All patients with 5–6 points sustained AVF. CONCLUSIONS More severe wedge angle before surgery, correction degree, old OVF presence, and thoracolumbar level were predictive factors for AVF. All patients with AVF risk score of 5 or more showed AVF. This information may aid preoperative risk assessment, informed shared decision-making, and consideration of potential alternative management strategies.